Space-Decomposition-Based Spectral Modeling for Distributed Battery Thermal Dynamics

The thermal behavior has sustainable impacts on the safety performance and cycle life of lithium-ion batteries. This article proposes a physics/data hybrid modeling framework for the thermal process of the pouch cell. Since the thermal effects of the tab area have different dynamic characteristics from other parts of the battery, a space decomposition strategy is designed to generate a more effective multiphysics-based model. To capture the battery thermal dynamics in real time, a reduced-order model is derived using the eigenfunction-based spectral expansion. All the unknown nonlinearities and other disturbances are further compensated through data-based learning. Simulation studies and experiments demonstrate the effectiveness of the proposed modeling methodology.

Automated summary

This article proposes a physics/data hybrid modeling framework for the thermal process of lithium-ion batteries. A space decomposition strategy is designed to generate a more effective model, and a reduced-order model and data-based learning are used to capture the battery's thermal dynamics. Simulation studies and experiments demonstrate the effectiveness of the proposed methodology.